Big Transactions on Galera Cluster

High Availability
22 September 3:10PM - 4:00PM @ Lausanne

50 minutes conference
Big and long term transactions have traditionally not behaved well in Galera Cluster. Large replication data sets have resulted in symptoms like elevated conflict rate, increased memory consumption, even OOM kill may happen, flow control pausing, cluster freezing... To deal with these issues, Galera users have been adviced to split large transactions in "reasonable size". There is also cluster configuration option for rejecting too large transactions, as a safe guard. This presentation explains the new large transaction replication method in Galera Cluster 4.0, which can now replicate transactions of any size, with minimal consequence for the cluster performance. And, as a nice side effect, how this new replication method can also be used to help in avoiding multi-master conflicts.


CEO, Codership
Database replication specialist, with long track record of developing various replication and clustering solutions for MySQL and PostgreSQL databases. Currently holding CEO position at Codership, working for Galera open source project with the target of developing a universally applicable synchronous multi-master replication solution.